Project Appraisal Guidelines - Railway Procurement Agency · 1.1. The planning and appraisal of any...

August 2012

Unit 16.2 Expansion Factors forShort Period Traffic Counts

Project Appraisal Guidelines

Page | 1

Project Appraisal Guidelines Unit 16.2

Expansion Factors for Short Period Traffic Counts

Version Date Comments

1.0 August 2012 New Guidance

This document is available to download at www.nra.ie/publications/projectappraisal

For further queries please contact:

Strategic Planning Unit

National Roads Authority

St Martin’s House

Waterloo Road

Dublin 4

Tel: (01) 660-2511

Email: [email protected]

Web: www.nra.ie

NRA Project Appraisal Guidelines Unit 16.2: Expansion Factors for Short Period Traffic Counts

Page | 7

Glossary

AADT Annual Average Daily Traffic

WADT Weekly Average Daily Traffic

AM Peak The hour in which traffic flow peaks in the morning period

Inter Peak Period The period between the AM Peak and the PM Peak periods

PM Peak Period The hour in which traffic flow peaks in the evening period

Definitions

Annual Average Daily Traffic (AADT)

The total volume passing a point or segment of a road for one year, divided by the

number of days in the year.

Weekday Average Daily Traffic (WADT)

The total volume passing a point or segment of a road for one full week (Monday to

Sunday), divided by the number of days in the week (7).

Average Weekday Traffic

The total volume passing a point or segment of a road for one full working week

(Monday to Friday), divided by the number of days in the working week (5).

Generic Expansion Factor Method

A method of extrapolating short period traffic counts to longer periods, or to other

count periods using a series of standard indices.

Traffic Flow Profile

The shape of a line plot of hourly traffic flow across a 24-hour period, with hourly

traffic flow expressed as a proportion of the total daily flow.


Page | 8

1. Overview

1.1. The planning and appraisal of any proposed works on the national road network

requires an understanding of the traffic demand that will be supported by such

improvements. This ensures that the proposed works can accommodate the

anticipated demand, and allows the quantification of benefits as part of the scheme

appraisal.

1.2. PAG Unit 16.1: Estimating AADT on National Roads sets out a methodology for

calculating AADT using short period traffic counts for input to scheme design and

appraisal. That Unit describes a series of options for additional data collection in

order to convert localised short period counts or traffic model outputs into AADT.

1.3. It is accepted, nevertheless, that at an early stage in scheme planning the collection

of data to support estimation of AADT may not be warranted. Instead, a means of

developing outline indications of AADT, or of understanding the variability in traffic

flow throughout the year is required. This allows the preparation of preliminary

analysis, which can be supported by more detailed data collection at a later stage in

scheme planning.

1.4. This PAG Unit has been prepared to support the conversion of Short Period Traffic

Counts to AADT, or to facilitate the estimation of short period traffic flows at any point

in a year using only a sample dataset of traffic information. The guidance achieves

this by developing an annual traffic flow profile that considers time of day, day of

week and month of year, thereby generating a series of indices to allow short period

counts from one period to be extrapolated to any period of the year, or to AADT.

This is referred to as the Generic Expansion Factor Method.

1.5. This guidance acknowledges the potential for statistical error in the Generic

Expansion Factor Method. Confidence limits are set out where appropriate to ensure

that the variability in observations can be considered.

1.6. The Generic Expansion Factor Method can be used for short period counts where

nearby data is not available, and where the collection of localised period counts is

not justified. This would normally be the case where data is required for high level

studies only, and where outputs will not be used as the basis for Project Appraisal or

any form of local capital investment. For all other applications, please refer to PAG

Unit 16.1: Estimating AADT on National Roads. The guidance presented in RT201 is

superseded by PAG Units 16.1 and 16.2.

1.7. For the purpose of this PAG Unit, short period traffic counts are defined as counts of

duration less than 7 days. The flowchart in Figure 16.2.1 should be used to inform

the most appropriate means of developing AADT estimates from short period traffic

counts in different circumstances.


Page | 9

Figure 16.2.1: Approaches for Estimating AADT from Short Period Counts

2. Factors Effecting Traffic Flow Profiles

2.1. The conversion of short period traffic counts to alternative short periods or to AADT

is dependant on the shape of the traffic flow profile across a defined period. Flatter

traffic flow profiles indicate a more uniform distribution of traffic flow across a

particular period, whilst roads with a more pronounced increase in traffic flow during

particular hours, days or months will lead to increased variability in the traffic flow

profile.

2.2. This is indicated below in Figure 16.2.2. The blue profile shows a distinct peak in

traffic flow during the AM and PM Peak periods, with very low flows overnight and a

notable reduction in flow during the middle of the day. The profile shown in red

shows a lower peak effect, with higher proportional flows during the middle of the day

and overnight. This demonstrates the impact of the daily flow profile on the validity

of expansion factors used to generate daily traffic flow estimates from short period

counts during the day.

NRA Project Appraisal Guidelines

Figure

2.3. In recognition of this, it is therefore necessary that this variability is considered in the

development of indices to be applied to short period traffic counts. In this regard, the

findings of PAG Unit 16.1: Estimating AADT on National Roads

which highlighted that the following factors influence the generation of indices for

expansion of short period traffic counts:

• Geographical Location;

• Day of Week; and

• Month of Year.

2.4. For the Generic Expansion Factor Method,

incorporated into the preparation of indices.

3. Overview of the Generic Expansion Factor Method

3.1. The conversion of short period traffic counts to alternative periods, and to facilitate

an outline estimation of AADT is based on the foll

• Hour of Day: A series of traffic flow profiles

weekdays for different geographical regions;

• Day of Week: Indices

to an average day for that week (WADT); and

• Month of Year: Indices

of AADT.

3.2. Although the use of these datasets has been presented to convert short period traffic

counts upwards to AADT, they can equally be


Figure 16.2.2: Daily Traffic Flow Profiles

In recognition of this, it is therefore necessary that this variability is considered in the


: Estimating AADT on National Roads are referen


expansion of short period traffic counts:

Geographical Location;

For the Generic Expansion Factor Method, the effects of these criteria will also be

the preparation of indices.

he Generic Expansion Factor Method

The conversion of short period traffic counts to alternative periods, and to facilitate

an outline estimation of AADT is based on the following datasets:

A series of traffic flow profiles that are representative of

ifferent geographical regions;

: Indices that allow the conversion of a single 24-hour traffic flow

an average day for that week (WADT); and

: Indices that allow the conversion of WADT into an estimate

Although the use of these datasets has been presented to convert short period traffic

counts upwards to AADT, they can equally be used to convert between different

Unit 16.2: Expansion Factors for Short Period Traffic Counts

Page | 10

In recognition of this, it is therefore necessary that this variability is considered in the


are referenced,


will also be

The conversion of short period traffic counts to alternative periods, and to facilitate

that are representative of

hour traffic flow

that allow the conversion of WADT into an estimate

Although the use of these datasets has been presented to convert short period traffic

used to convert between different


Page | 11

short periods thorough calculation of AADT, followed by inverted use of indices to

estimate traffic flow for any short period.

4. Factoring by Hour of Day

4.1. 24-hour traffic flow profiles have been developed based on a series of geographical

regions. In order to define regions, it was necessary to develop an indicator of the

shape of the flow profile over a typical day.

4.2. The shape of a daily flow profile can either be described as ‘peaky’ or ‘flat’. Peaky

flow profiles exhibit high traffic volumes during AM and PM Peak periods, with

relatively lower flow throughout the remainder of the day (shown as blue in Figure

16.2.2. This pattern is most evident on busy commuting corridors, with a limited

number of non-employment related uses to generate ongoing demand throughout

the day.

4.3. Flat profiles exist for a number of reasons. Congested roads can exhibit relatively

flat profiles. The existence of a capacity constraint during peak periods can lead to

spreading of peak demand into the peak shoulders, effectively restraining the peak

flow whilst increasing the off-peak flows. In the most severe cases, the existence of

bottlenecks or opposing movements downstream can lead to a significant reduction

in traffic flow during the peak periods. This can skew the results of any expansion of

peak hour flows quite considerably, and supports the conclusion that the use of short

period traffic flows during congested periods should not be used as the basis for any

estimation of WADT or AADT.

4.4. Nevertheless, flat profiles may also suggest that there is a broader range of trip

purposes along a corridor, with commuting demand during the peak periods overlaid

with retail, leisure or other business demands during the off peak periods. As such,

flatter flow profiles can easily exist where there are no significant capacity limitations.

4.5. Finally, the use of flow profiles is not appropriate on routes where variable price

tolling is employed (such as the Dublin Port Tunnel). In such cases, the variable

pricing of tolls influences demand and traffic volumes can deviate significantly from

the traffic flow profiles presented here.

4.6. In order to represent the ‘Peakiness’ of a traffic flow profile over a particular day, the

concept of a ‘p-factor’ has been derived. The p-factor simply describes the scale of

the reduction in flow between the AM Peak and the quietest period of the afternoon

(the Inter-Peak), and from the Inter-Peak back up to the PM Peak. It is defined as

follows:

� � � � � � 2�

where: p = the peak factor (known as p-factor) a = the maximum hourly proportion of traffic between 0:00 and 12:00 on a weekday b = the maximum hourly proportion of traffic between 12:00 and 24:00 on a weekday c = the minimum hourly proportion of traffic between 08:00 and 18:00 on a weekday


Page | 12

4.7. Consider the example below in Figure 16.2.3. The values for a, b, and c are shown

on the graph. The p-factor yields a value of 0.057. Note that the maximum p-factor

is 1.0, in which case all traffic flow would occur during 2 individual peak hours of the

day, separated by a cessation of all traffic during the afternoon. The minimum

possible theoretical value is zero.

Figure 16.2.3: Calculation of p-factor

4.8. In order to understand the impact of geographical location on the p-factor, and hence

the traffic flow profile, p-factors for NRA Permanent Counters have been calculated.

Figure 16.2.4 presents a frequency plot of values over the range of p-factors from 0

to 0.15.

4.9. The analysis of p-factors demonstrates that the range of data is bounded by an

upper 85 percentile value of 0.081, with a lower 85 percentile value of 0.043. The

analysis also shows a cluster of values around the mean of 0.062.


Page | 13

Figure 16.2.4: Distribution of p-factor at NRA Permanent Count Sites

4.10. As a further exercise, the range of p-factors has been plotted thematically to

understand the influence of geography. This is presented in Figure 16.2.5 and

shows high p-factor values as red/yellow, with lower values in blue. Mid range

values are shown as green.


Page | 14

Figure 16.2.5: Thematic Mapping of p-factors at NRA Permanent Count Sites

4.11. The mapping shows a definite influence of geography on the p-factor, and hence on

the shape of the traffic flow profile. There is a pattern of high p-factor values in the

west and southwest, and to a lesser extent through the Greater Dublin Area. The

lowest values are evident through the midlands.


Page | 15

4.12. On closer examination of individual results, it becomes clear that the Major Inter-

Urban Routes exhibit relatively low p-factor values across the network. This

suggests that the pattern of activity on the MIU’s is substantively different to other

roads, and warrants a separate classification.

4.13. As a result of the above findings, it is appropriate to develop a range of traffic flow

profiles which reflect the different patterns across the Country, effectively reducing

the standard deviation in p-factors within each area and facilitating a more accurate

estimation of the 24-hour traffic flow from a short period count. The following

geographies have been identified:

• The Major Inter-Urbans (excluding the Greater Dublin Area), which exhibit the

lowest p-factors;

• The Southwest and West, comprising Counties Cork, Kerry, Limerick, Clare,

Galway and Mayo;

• The Greater Dublin Area; and

• All other areas.

4.14. The relevant Traffic Flow Profiles for these areas are outlined in Annex A of this PAG

Unit. A series of examples outlining the conversion of short period traffic counts to

24-hour weekday totals are outlined overleaf.

4.15. Confidence Intervals for expansion factors for each hour to the 24-hour total are also

presented in Annex A based on a 95% Confidence Interval. The table shows those

values which yield intervals greater than +20% in grey, suggesting that use of single-

hour data from these periods will not yield a reliable result. For the estimation of 24-

hour flow for traffic counts from these periods, they should be combined with flows

from other periods to improve the accuracy of the estimation. In the case of counts

from multiple periods, the 95% Confidence Interval may be assumed as the lowest

value of the various hours for which data is available.


Page | 16

Example: Conversion of Short Period Traffic Count to a 24-hour total.

Consider a traffic count covering the period from 08:00 to 09:00 on the N5 at Bohola,

County Mayo. The traffic count reports a traffic flow of 600 vehicles per hour (2-

way).

Step 1: Identify the relevant geographical location. For this example, the

relevant profiles are for the Southwest and West.

Step 2: Identify the proportion of daily traffic flow that relates to the short

period count. From Annex A it can be seen that the proportion of the

24-hour flow occurring during the period 08:00 – 09:00 is 0.073.

Step 3: Perform the calculation. The 24-hour estimate is therefore as follows

600/0.073 = 8219

The 24-hour estimate is therefore 8219 vehicles

Example: Conversion of Short Period Traffic Count to a 24-hour total.

Consider a traffic count covering the period from 08:00 to 09:00 and 14:00 – 16:00

on the N11 at Ferrycarraig, Co Wexford. The traffic count reports a traffic flow of

1400 vehicles (2-way) during the combined hours.

Step 1: Identify the relevant geographical location. For this example, the

relevant profiles are for All Other Areas.

Step 2: Identify the proportion of daily traffic flow that relates to the short

period count. From Annex A it can be seen that the proportion of the

24-hour flow occurring during the period 08:00 – 09:00 and 14:00 –

16:00 is 0.215. (This is calculated as the sum of 0.07, 0.07 and 0.075

which cover the three individual time periods.

Step 3: Perform the calculation. The 24-hour estimate is therefore as follows

1400/0.215 = 6512

The 24-hour estimate is therefore 6512 vehicles


Page | 17

5. Factoring by Day of Week

5.1. Once a 24-hour estimate is derived, a further calculation can be undertaken to

convert this to an average weekday (or to estimate 24-hour flows that might arise on

other weekdays).

5.2. Figure 16.2.5 below shows a weekly flow profile based on data from the network of

NRA Permanent Traffic Counters. The data shows a gradual increase in the 24-hour

traffic flow from Monday through to a peak on Friday, with a reduction in total flows at

weekends.

Figure 16.2.5: Weekly Flow Profile

5.3. Note that Weekly Average Daily Traffic (WADT) refers to the mean 24-hour traffic

flow across a week. This is not to be confused with a Weekday Average, which

represents a 24-hour average for the Monday to Friday period. Geography is not

considered to be a significant factor in converting from a 24-hour estimate to WADT.

5.4. A series of Weekly Flow Indices have been generated using this data to enable the

conversion from a specific day to a Weekly Average day. These are outlined in

Annex B of this PAG Unit. An example of a conversion from a single day to an

estimate of the Weekly Average Daily Traffic is outlined below.


Page | 18

Example: Conversion of 24-hour flow to a Weekly Average Day.

Consider a 24-hour traffic volume collected on a Tuesday. The traffic count reports a

traffic flow of 12,500 vehicles over the 24-hours.

Step 1: Identify the relevant day, which in this case is a Tuesday.

Step 2: Extract the correct factor to convert to Weekday Average Day. For a

Tuesday, Annex B shows that the relevant factor is 0.98

Step 3: Perform the calculation. The WADT is therefore as follows

12,500*0.98 = 12,250

The WADT estimate is therefore 12,250 vehicles

6. Factoring by Month of Year

6.1. The final calculation is to convert WADT to AADT. This final stage recognises the

significant variation that can occur in traffic flows across the year. Such variations

arise as a result of:

• The seasonal effects of tourism;

• Reductions in school trips during holiday periods;

• The effect of annual leave on the volume of commuting trips during the

summer months; and

• Changes in the level of retail activity.

6.2. This effect is also referred to as seasonality, which describes the variability in traffic

flow at different times of the year. To demonstrate the scale of the seasonality effect,

Figure 16.2.6 shows WADT for each month expressed as a proportion of AADT. The

data shows that daily traffic flows peak during the summer period, with the lowest

flows during the months of December and January.


Page | 19

Figure 16.2.6: Monthly Flow Profile

6.3. It is noted that the relatively low level of traffic flow during December and January is

influenced by the Christmas period where traffic flows reduce significantly over a

period of 2 weeks. This therefore reduces WADT for that period. Likewise, the high

levels of traffic in August are driven by the increased levels of tourism and leisure

movements which tend to take place outside peak periods.

6.4. For the purpose of this PAG Unit, the monthly indices for December and January

have been adjusted to remove the influence of the Christmas Period. In this regard,

these relationships do not hold for short period traffic counts undertaken during

periods of abnormal traffic flow (i.e. Christmas, Easter or Bank Holiday weekends).

Furthermore, care should be taken when using short period traffic counts for July and

August to extrapolate to AADT due to the influence of tourist traffic on the 24-hour

traffic flow profile. Where possible, some conversion to 24-hour data using NRA

Permanent Counters or Temporary Count Sites should be considered.

6.5. Monthly Flow Indices to account for the influence of seasonality are outlined in

Annex C of this PAG Unit. An example of a conversion from a single day to an

estimate of the Weekly Average Daily Traffic is outlined below.


Page | 20

Example: Conversion of Weekly Average Day to Annual Average Day.

Consider a Weekly Average Daily Traffic Flow of 10,000 vehicles on a site in Co.

Roscommon during March. This is to be converted to an estimate of AADT.

Step 1: Identify the relevant month, which in this case is March.

Step 2: Extract the correct factor to convert to Annual Average Day. For the

Month of March, Annex C shows that the relevant factor is 1.05.

Step 3: Perform the calculation. The AADT is therefore as follows

10,000*1.05 = 10,500.

The AADT estimate is therefore 10,500 vehicles

7. Dealing with Multiple Datasets

7.1. In some cases, a selection of data may be available to support the estimation of

AADT. For example, a particular location may provide data for March and a

separate dataset for May, and estimating AADT from these datasets individually will

provide different results.

7.2. Whilst some variability in the results is expected, it is noted that the procedure

outlined in this PAG Unit is intended to give no more than an indicative estimate of

AADT based on short period counts. Where more detailed estimates are required,

reference should be made to PAG Unit 16.1: Estimating AADT on National Roads.

7.3. Where multiple results for a single site can be derived based on alternative data, the

user should take a practical view as to the conclusions that can be inferred from

them. Obviously the more data that is captured within the short period traffic count;

the more robust will be the final calculation.


Page | 21

Annex A

Traffic Flow Profiles


Page | 22


Page | 23


Page | 24


Page | 25


Page | 26

Hourly Traffic Flow as a Proportion of 24-hour Traffic Flow Hour Ending 01:00 02:00 03:00 04:00 05:00 06:00 07:00 08:00 09:00 10:00 11:00 12:00

Greater Dublin Area 0.008 0.005 0.003 0.004 0.006 0.017 0.049 0.072 0.073 0.059 0.053 0.054

Major Inter-Urbans (Excl. GDA) 0.009 0.005 0.004 0.004 0.006 0.013 0.030 0.064 0.077 0.062 0.055 0.057

West and Southwest 0.007 0.004 0.002 0.002 0.003 0.006 0.021 0.062 0.073 0.060 0.056 0.059

All Other Areas 0.007 0.004 0.003 0.003 0.004 0.011 0.026 0.060 0.070 0.061 0.056 0.059

Hour Ending 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 24:00

Greater Dublin Area 0.056 0.058 0.061 0.068 0.079 0.085 0.074 0.056 0.041 0.032 0.023 0.014

Major Inter-Urbans (Excl. GDA) 0.060 0.063 0.065 0.071 0.082 0.090 0.074 0.054 0.041 0.031 0.022 0.014

West and Southwest 0.062 0.063 0.065 0.071 0.081 0.089 0.075 0.054 0.039 0.030 0.021 0.013

All Other Areas 0.064 0.067 0.070 0.075 0.084 0.090 0.073 0.055 0.041 0.031 0.021 0.012

95% Confidence Intervals for 24-hour Flow Estimations

Hour Ending 01:00 02:00 03:00 04:00 05:00 06:00 07:00 08:00 09:00 10:00 11:00 12:00

Greater Dublin Area +58% +62% +62% +51% +77% +103% +55% +33% +23% +15% +12% +13%

Major Inter-Urbans (Excl. GDA) +41% +52% +45% +36% +60% +63% +50% +36% +26% +17% +11% +8%

%West and Southwest +42% +54% +60% +68% +91% +79% +66% +62% +38% +13% +17% +21%

All Other Areas +64% +67% +72% +73% +100% +83% +67% +54% +24% +15% +12% +11%

Hour Ending 13:00 14:00 15:00 16:00 17:00 18:00 19:00 20:00 21:00 22:00 23:00 24:00

Greater Dublin Area +13% +12% +11% +13% +14% +17% +18% +22% +28% +31% +39% +46%

Major Inter-Urbans (Excl. GDA) +10% +11% +9% +10% +15% +17% +13% +16% +21% +18% +24% +31%

West and Southwest +17% +13% +14% +13% +13% +17% +11% +16% +24% +25% +28% +35%

All Other Areas +13% +13% +13% +16% +18% +17% +20% +27% +27% +31% +38% +51%


Page | 27

Annex B

Weekly Flow Indices


Page | 28

Weekly Flow Indices – All Locations Day Mon Tue Wed Thu Fri Sat Sun

Proportion of WADT 0.99 1.02 1.03 1.07 1.15 0.91 0.86

Index 1.01 0.98 0.97 0.93 0.87 1.10 1.16

Note: Proportion of WADT measures the level of the 24-hour flow for any day against the weekly average day (i.e. traffic flow for a Tuesday is 1.02 times the Weekday Average Daily Traffic). The Index is the factor that is applied to the 24-hour total to convert it to WADT.


Page | 29

Annex C

Monthly Flow Indices


Page | 30

Monthly Flow Indices – All Locations Day Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Proportion of AADT 0.94 0.94 0.96 1.02 1.03 1.04 1.06 1.08 1.04 1.01 0.96 0.96

Index 1.06 1.06 1.04 0.98 0.97 0.96 0.94 0.93 0.96 0.99 1.04 1.04

Note: Proportion of AADT measures the level of the WADT for each month against AADT (i.e. WADT in March is 0.96 times AADT). The Index is the factor that is applied to WADT to convert it to AADT.

Project Appraisal Guidelines - Railway Procurement Agency · 1.1. The planning and appraisal of any...

Documents

Transcript of Project Appraisal Guidelines - Railway Procurement Agency · 1.1. The planning and appraisal of any...